On Wed, Jan 08, 2025 at 01:21:05PM +0100, m...@asdfghasdfgh.de wrote: > Hi! > > I have a KVM VPS with IN-Berlin.de, running OpenBSD 7.6 and around every > second day of uptime, OpenBSD crashes. How shall I proceed, debugging > this, to eventually report a bug? Please see the attached dmesg. The > swap is 753M, the processes are usually as listed in the attached file. > > qemu-ga is installed and ddb.panic=0 set. One problem: if the crash > happens, I can’t access the serial console of the VPS anymore, which > is accessible via SSH as com0. This means I can’t run ddb… > > An alike OpenBSD KVM VPS runs without problems, albeit with another > hosting company. >
> panic: pool_do_get: mcl2k free list modified: page 0xfffffd801eb7a000; item > addr 0xfffffd801eb7c800; offset 0x0=0x0 != 0x8bfa738a7ce007d4 > Starting stack trace... > panic(ffffffff82338b22) at panic+0x167 > pool_do_get(ffffffff82872690,2,ffff80001ca65d74) at pool_do_get+0x350 > pool_get(ffffffff82872690,2) at pool_get+0x86 > m_clget(0,2,800) at m_clget+0x1a8 > vio_populate_rx_mbufs(ffff80000008a800,ffff8000000ab300) at > vio_populate_rx_mbufs+0x136 > vio_rx_intr(ffff800000115000) at vio_rx_intr+0x73 > intr_handler(ffff80001ca65f20,ffff8000000ac000) at intr_handler+0x56 > Xintr_ioapic_edge22_untramp() at Xintr_ioapic_edge22_untramp+0x18f > memset() at memset+0x5c > wg_encap_worker(ffff8000009df000) at wg_encap_worker+0x8b > taskq_thread(ffff8000000b2800) at taskq_thread+0x129 > end trace frame: 0x0, count: 246 > End of stack trace. This is most probably solved in -current. At least no new reports have been sent to the list about wg(4) crashes after rev. 1.39 commit of sys/net/if_wg.c So maybe try -current -- :wq Claudio
